Public bug reported:
The 5.8 kernel series in Groovy (20.10) drops the video signal to
DisplayPort output on my AMD Radeon HD 7950 card. GRUB comes up, but the
signal is dropped when switching to the framebuffer (I believe).
Falling back to the 5.4 kernel series works. So does nomodeset.
However, a LiveUSB does not. Nor do any of the 5.6 to 5.10 kernels.
With no video output, ssh-ing into the dark machine generates the
following:
GPU specs:
[code]
duckhook@Zeus:~$ sudo lspci -vk | grep -iA20 vga
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI]
Tahiti PRO [Radeon HD 7950/8950 OEM / R9 280] (prog-if 00 [VGA controller])
Subsystem: Hightech Information System Ltd. Tahiti PRO [Radeon HD 7950/8950
OEM / R9 280]
Flags: bus master, fast devsel, latency 0, IRQ 56
Memory at d0000000 (64-bit, prefetchable) [size=256M]
Memory at fbe00000 (64-bit, non-prefetchable) [size=256K]
I/O ports at e000 [size=256]
Expansion ROM at 000c0000 [disabled] [size=128K]
Capabilities: [48] Vendor Specific Information: Len=08 <?>
Capabilities: [50] Power Management version 3
Capabilities: [58] Express Legacy Endpoint, MSI 00
Capabilities: [a0] MSI: Enable+ Count=1/1 Maskable- 64bit+
Capabilities: [100] Vendor Specific Information: ID=0001 Rev=1 Len=010 <?>
Capabilities: [150] Advanced Error Reporting
Capabilities: [270] Secondary PCI Express
Capabilities: [2b0] Address Translation Service (ATS)
Capabilities: [2c0] Page Request Interface (PRI)
Capabilities: [2d0] Process Address Space ID (PASID)
Kernel driver in use: radeon
Kernel modules: radeon, amdgpu
[/code]
A comparison between a working session under 5.4 vs a dark session under 5.8
yields the difference of these two lines in journalctl:
[code]
duckhook@Zeus:~$ sudo journalctl -b | grep -i radeon | grep kernel
…
Nov 15 18:49:04 Zeus kernel: [drm:radeon_dp_link_train_cr [radeon]] *ERROR*
clock recovery tried 5 times
Nov 15 18:49:04 Zeus kernel: [drm:radeon_dp_link_train_cr [radeon]] *ERROR*
clock recovery failed
[/code]
…
The full log is attached.
Through mainline kernel bisecting, I have traced the problem to the following
build inflection point:
Kernel 5.5.5 → Good
Kernel 5.5.6 → Bad
This is pretty much the limit of my technical capabilities. I am not
knowledgeable enough to produce an actual commit kernel bisection, but
playing a layman's hunch (and at the real risk of leading devs astray),
the only reference to radeon in the 5.5.6 changelog is the following:
Daniel Vetter (1):
radeon: insert 10ms sleep in dce5_crtc_load_lut
Clearly, there are many other changes in this changelog dealing with the
video subsystem that I am not qualified to comment on.
Please ask if further info is needed. I am happy to supply all that is
within my technical capabilities.
** Affects: linux-hwe-5.8 (Ubuntu)
Importance: Undecided
Status: New
** Attachment added: "journalctl.txt"
https://bugs.launchpad.net/bugs/1904647/+attachment/5435339/+files/journalctl.txt
** Description changed:
The 5.8 kernel series in Groovy (20.10) drops the video signal to
DisplayPort output on my AMD Radeon HD 7950 card. GRUB comes up, but the
signal is dropped when switching to the framebuffer (I believe).
Falling back to the 5.4 kernel series works. So does nomodeset.
However, a LiveUSB does not. Nor do any of the 5.6 to 5.10 kernels.
With no video output, ssh-ing into the dark machine generates the
following:
GPU specs:
[code]
duckhook@Zeus:~$ sudo lspci -vk | grep -iA20 vga
-
+
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI]
Tahiti PRO [Radeon HD 7950/8950 OEM / R9 280] (prog-if 00 [VGA controller])
- Subsystem: Hightech Information System Ltd. Tahiti PRO [Radeon HD
7950/8950 OEM / R9 280]
- Flags: bus master, fast devsel, latency 0, IRQ 56
- Memory at d0000000 (64-bit, prefetchable) [size=256M]
- Memory at fbe00000 (64-bit, non-prefetchable) [size=256K]
- I/O ports at e000 [size=256]
- Expansion ROM at 000c0000 [disabled] [size=128K]
- Capabilities: [48] Vendor Specific Information: Len=08 <?>
- Capabilities: [50] Power Management version 3
- Capabilities: [58] Express Legacy Endpoint, MSI 00
- Capabilities: [a0] MSI: Enable+ Count=1/1 Maskable- 64bit+
- Capabilities: [100] Vendor Specific Information: ID=0001 Rev=1 Len=010
<?>
- Capabilities: [150] Advanced Error Reporting
- Capabilities: [270] Secondary PCI Express
- Capabilities: [2b0] Address Translation Service (ATS)
- Capabilities: [2c0] Page Request Interface (PRI)
- Capabilities: [2d0] Process Address Space ID (PASID)
- Kernel driver in use: radeon
- Kernel modules: radeon, amdgpu
+ Subsystem: Hightech Information System Ltd. Tahiti PRO [Radeon HD 7950/8950
OEM / R9 280]
+ Flags: bus master, fast devsel, latency 0, IRQ 56
+ Memory at d0000000 (64-bit, prefetchable) [size=256M]
+ Memory at fbe00000 (64-bit, non-prefetchable) [size=256K]
+ I/O ports at e000 [size=256]
+ Expansion ROM at 000c0000 [disabled] [size=128K]
+ Capabilities: [48] Vendor Specific Information: Len=08 <?>
+ Capabilities: [50] Power Management version 3
+ Capabilities: [58] Express Legacy Endpoint, MSI 00
+ Capabilities: [a0] MSI: Enable+ Count=1/1 Maskable- 64bit+
+ Capabilities: [100] Vendor Specific Information: ID=0001 Rev=1 Len=010 <?>
+ Capabilities: [150] Advanced Error Reporting
+ Capabilities: [270] Secondary PCI Express
+ Capabilities: [2b0] Address Translation Service (ATS)
+ Capabilities: [2c0] Page Request Interface (PRI)
+ Capabilities: [2d0] Process Address Space ID (PASID)
+ Kernel driver in use: radeon
+ Kernel modules: radeon, amdgpu
[/code]
A comparison between a working session under 5.4 vs a dark session under 5.8
yields the difference of these two lines in journalctl:
[code]
duckhook@Zeus:~$ sudo journalctl -b | grep -i radeon | grep kernel
…
Nov 15 18:49:04 Zeus kernel: [drm:radeon_dp_link_train_cr [radeon]] *ERROR*
clock recovery tried 5 times
Nov 15 18:49:04 Zeus kernel: [drm:radeon_dp_link_train_cr [radeon]] *ERROR*
clock recovery failed
[/code]
…
The full log is attached.
Through mainline kernel bisecting, I have traced the problem to the following
build inflection point:
Kernel 5.5.5 → Good
Kernel 5.5.6 → Bad
This is pretty much the limit of my technical capabilities. I am not
knowledgeable enough to produce an actual commit kernel bisection, but
playing a layman's hunch (and at the real risk of leading devs astray),
the only reference to radeon in the 5.5.6 changelog is the following:
Daniel Vetter (1):
- radeon: insert 10ms sleep in dce5_crtc_load_lut
+ radeon: insert 10ms sleep in dce5_crtc_load_lut
Clearly, there are many other changes in this changelog dealing with the
video subsystem that I am not qualified to comment on.
+
+ Please ask if further info is needed. I am happy to supply all that is
+ within my technical capabilities.
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1904647
Title:
Radeon driver drops DP signal when switching to framebuffer in 5.8
kernel
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux-hwe-5.8/+bug/1904647/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
